Elevated Inflammatory Response in Caveolin-1-deficient Mice with Pseudomonas aeruginosa Infection Is Mediated by STAT3 Protein and Nuclear Factor κB (NF-κB)*

Caveolin-1 (Cav-1), an important composition protein within the flask-shaped membrane invaginations termed caveolae, may play a role in host defense against infections. However, the phenotype in Pseudomonas aeruginosa-infected cav1 knock-out (KO) mice is still unresolved, and the mechanism involved is almost entirely unknown. Using a respiratory infection model, we confirmed a crucial role played by Cav-1 in host defense against this pathogen because Cav-1 KO mice showed increased mortality, severe lung injury, and systemic dissemination as compared with wild-type (WT) littermates. In addition, cav1 KO mice exhibited elevated inflammatory cytokines (IL-6, TNF-α, and IL-12a), decreased phagocytic ability of macrophages, and increased superoxide release in the lung, liver, and kidney. We further studied relevant cellular signaling processes and found that STAT3 and NF-κB are markedly activated. Our data revealed that the Cav-1/STAT3/NF-κB axis is responsible for a dysregulated cytokine response, which contributes to increased mortality and disease progression. Moreover, down-regulating Cav-1 in cell culture with a dominant negative strategy demonstrated that STAT3 activation was essential for the translocation of NF-κB into the nucleus, confirming the observations from cav1 KO mice. Collectively, our studies indicate that Cav-1 is critical for inflammatory responses regulating the STAT3/NF-κB pathway and thereby impacting P. aeruginosa infection.

(SOCSs) function as negative regulators of STATs and operate by binding and inhibiting JAKs or competing with STATs for phosphotyrosine binding sites on cytokine receptors (25). Recently, STAT3 was found to interact with Cav-1 and heat shock protein 90 in plasma membrane rafts during Escherichia coli infection (26). Cav-1 is also related to a JAK2/STAT5 pathway because Cav-1 is homologous to the pseudosubstrate for SOCS. Thus, Cav-1 may down-regulate JAK/STAT pathways, modulating the proinflammatory response. Because Cav-1 was found to be associated with STAT3 in lipid rafts by a co-localization study, it is also possible that a Cav-1 cascade may impact the PI3K/Akt pathway through molecular interactions, which may serve as a regulator for STAT3 (27). Thus, Cav-1 may help maintain the balance between host response and tissue damage that may be caused by overproduction of inflammatory cytokines.
To better characterize the role of Cav-1 in P. aeruginosa infection, we sought to determine the pathogenic mechanism in cav1 KO mice. We found that cav1 KO mice manifested more severe infection, including increased bacterial burdens, inflammation, oxidative stress, and susceptibility to infection. Our studies also showed that the STAT3/NF-B pathway is responsible for the dysregulated response during P. aeruginosa infection.

EXPERIMENTAL PROCEDURES
Mouse cav1-KO and wild-type (WT) control mice (B6129SF2/J) were obtained from The Jackson Laboratory (Bar Harbor, ME) (28). Mice were housed and bred in the animal facility at the University of North Dakota, and the animal experiments were performed in accordance with the institutional animal care and use committee guidelines. We anesthetized mice with 45 mg/kg ketamine and intranasally instilled 0.5 ϫ 10 7 (PAK; six mice/group) to 1 ϫ 10 7 (PAO1) colony-forming units (cfu) of P. aeruginosa, and mice were killed when they were moribund (29). After bronchoalveolar lavage (BAL), the trachea and lung were excised for homogenization or fixed in 10% formalin.
Cells-Mouse AM cells were isolated by BAL as described (30). AM cells were grown in RPMI 1640 medium supplemented with 10% newborn calf serum and penicillin/streptomycin antibiotics in a 5% CO 2 incubator. MLE-12 cells were obtained from ATCC and maintained following the manufacturer's instructions.
Infection Experiments-Bacteria were grown overnight in Luria-Bertani (LB) broth at 37°C with vigorous shaking. The next day, the bacteria were pelleted by centrifugation at 5,000 ϫ g, resuspended in 10 ml of fresh LB broth, and allowed to grow until the midlogarithmic phase. Optical density (OD) at 600 nm was measured, and density was adjusted to 653,740.25 OD (0.1 OD ϭ 1 ϫ 10 8 cells/ml). Cells were washed once with PBS after overnight culture in serum-containing medium and changed to serum-free and antibiotic-free medium immediately before infection (29). Except for dose determination assays, cells were infected by P. aeruginosa in a multiplicity of infection (m.o.i.) of 10:1 bacteria-cell ratio for the indicated time points.
Inflammatory Cytokine Profiling-Cytokine concentrations were measured by an ELISA kit (eBioscience Co., San Diego, CA) in samples of cell culture medium, BAL fluid, and lung homogenates collected at the indicated times after infection. The MLE-12 cells were treated as described above. Culture medium was collected after infection. For BAL fluid, the trachea was surgically exposed and cannulated, lungs were lavaged five times with 1.0-ml volume of lavage fluid, the lavageates were pooled, and cells were removed by centrifugation. For lung homogenates, excised lungs were ground in 500 l of PBS. 96-well plates (Corning Costar 9018) were coated with 100 l/well capture antibody in coating buffer and incubated overnight at 4°C (33). 100-l aliquots of serum samples were added to the coated microtiter wells. The cytokine concentrations were determined with corresponding detection HRP-conjugated antibodies. The values were read at 450 nm and analyzed.
Western Blotting-Mouse monoclonal Abs against Cav-1, phospho-STAT3, IL-6, and NF-B; rabbit polyclonal Ab against phospho-NF-B; and goat polyclonal Abs against TNF-␣ and SOCS3 were from Santa Cruz Biotechnology, Inc. Rabbit monoclonal Ab against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was from Cell Signaling Technology. The samples derived from cells and lung homogenates were lysed and quantitated. The lysates were boiled for 5 min, and protease inhibitor mixture was added. The supernatants were collected, and an equal amount of each sample was loaded onto 10% SDS-polyacrylamide minigels and electrophoresed to resolve proteins. The proteins were then transferred to polyvinylidene difluoride membranes (Pierce) and blocked for 2 h at room temperature using 5% nonfat milk blocking buffer. Membranes were incubated overnight at 4°C with appropriate first antibodies diluted at 1:1,000 in 5% bovine serum albumin (BSA) Western antibody buffer. After washing three times with washing solution, the antigen-antibody complexes were incubated for 45 min at room temperature with horseradish peroxidase-conjugated secondary antibody (Rockland Immunochemicals, Gilbertsville, PA) diluted 1:2,000 (34). Signals were visualized using an enhanced chemiluminescence detection kit (SuperSignal West Pico, Pierce).
Reverse Transcription (RT)-PCR Analysis-RNA was extracted from lung homogenates and cells with TRIzol (Invitrogen) according to the manufacturer's instructions. For detected genes, RT was performed using 1.5 g of RNA, RNase ribonuclease inhibitor, oligo(dT), and cloned avian myeloblastosis virus reverse transcriptase (Invitrogen). PCR products were separated by 1.0% agarose gel electrophoresis containing ethidium bromide and visualized under UV light. The results for each gene were normalized in comparison with GAPDH expression (35).

Confocal Microscopy and Indirect Immunofluorescence
Staining-Cells were grown either on coverslips in a 24-well plate or in glass bottom dishes (MatTek, Ashland, MA). For immunostaining, the cells were fixed in 3.7% paraformaldehyde and permeabilized with 0.2% Triton X-100 in PBS, and the nonspecific binding site was blocked with blocking buffer for 30 min. Cells were incubated with primary Abs at a 1:500 dilution in blocking buffer for 1 h and washed three times with wash buffer. After incubation with appropriate fluorophore-conjugated secondary Abs, the coverslips were mounted on slides with VECTASHIELD mounting medium (36). DAPI (Sigma-Aldrich) was used to stain the nucleus. The images were captured by an LSM 510 Meta confocal microscope (Carl Zeiss MicroImaging, Thornwood, NY) and processed using the software provided by the manufacturer.
Luciferase Assay-Transient transfections were performed with 2 ϫ 10 5 cells plated in 6-well plates by using 2 g of DNA and 3 l of Lipofectamine 2000 reagent (Invitrogen) in serumfree RPMI 1640 medium following the manufacturer's instruction. 24 h after transfection, the cells were infected with PAK at 10:1 m.o.i. for 1 h (37). Cell lysates were subjected to luciferase activity analysis by using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI).
3-(4,5-Dimethylthiazol-2-yl)-2,5-dimethyltetrazolium Bromide Assay-This assay measures the color change of 3-(4,5dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide upon reduction by enzymes to assess the viability of cells. BAL fluid was centrifuged, and cells were cultured in a 96-well plate. After incubation with P. aeruginosa, 3-(4,5-dimethylthiazol-2-yl)-2,5-dimethyltetrazolium bromide dye was added at a final concentration of 1 g/ml per well. The cells were incubated at 37°C for 60 min or until the color change occurred. The dye is yellow in color and upon reduction by enzymes forms a blue formazan product. The reaction was stopped by adding 100 l of stop solution (10% DMSO, 10% SDS in 50 mM HEPES buffer). The plate was left at room temperature overnight for complete dissolution of formazan crystals. The next day, the plate was read at 560-nm absorbance using a plate reader to quantify the dye conversion (38). Duplicates were done for each sample and control. Background correction was done with blanks containing dye alone.
NBT Assay-This assay is based on the color change of NBT dye upon reduction by released superoxide. Cells were treated as above, and same amount of dye as used above was added. The dye is yellow in color and upon reduction by superoxide forms a blue formazan product (39).
Dihydrodichlorofluorescein Diacetate Assay-Dihydrodichlorofluorescein diacetate dye (Molecular Probes, Carlsbad, CA) does not normally fluoresce but emits green fluorescence upon reaction with superoxide inside cells. Cells were treated as above, and an equal amount(s) of dye was added. After a 10-min incubation, fluorescence was measured using a fluorometer (BioTek, Winooski, VT) (39,40).
Lipid Peroxidation Assay-Malondialdehyde is an end product of the lipid peroxidation process and was measured in a colorimetric assay (Calbiochem) according to the manufacturer's instructions. Homogenized lung tissue in 62.5 mM Tris-HCl (pH ϭ 6.8) supplemented with Complete Mini protease inhib-itor (Roche Diagnostics) in equal protein amounts was used in the assay. Duplicates were done for each sample and control (41).
Phagocytosis Assay-AM cells obtained from BAL fluid were plated in 96-well plates and grown overnight. The cells were treated with the serum-free medium for 1 h. Then GFP-PAO1 was used to infect the cells at 10:1 m.o.i. After a 1-h incubation at 37°C, the wells were washed and treated with 100 g/ml polymyxin B for 1 h to kill any remaining extracellular bacteria (29). The number of phagocytosed bacteria was counted using a Synergy HT fluorometer (BioTek) with 485 Ϯ 20-nm excitation and 528 Ϯ 20-nm emission filters. Background correction was done for autofluorescence. Duplicates were done for each sample and control. Alternatively, classical colony-forming units were counted to quantitate phagocytosis as described previously.
Myeloperoxidase (MPO) Assay-The MPO assay was performed as follows. Samples were homogenized in 50 mM hexadecyltrimethylammonium bromide, 50 mM KH 2 PO4, pH 6.0, 0.5 mM EDTA at 1 ml/100 mg of tissue and centrifuged for 15 min at 12,000 rpm at 4°C. Supernatants were decanted, and 100 ml of reaction buffer (0.167 mg/ml O-dianisidine, 50 mM KH 2 PO 4 , pH 6.0, 0.0005% mM H 2 O 2 ) were added to 100 ml of sample. Absorbance was read at 460 nm at 2-min intervals. Duplicates were done for each sample and control (41).
Bacterial Burden Assay-AM cells from BAL fluid and ground lung, spleen, liver, and kidney tissues were homogenized with PBS and spread on LB plates to enumerate levels of bacteria. The plates were cultured in a 37°C incubator overnight, and colonies were counted. Duplicates were done for each sample and control.
Histopathology Analysis-Lung tissues were fixed in 10% formalin and then embedded in paraffin using a routine histologic procedure. Four-micrometer sections were cut, stained by standard H&E, and examined for differences in morphology postinfection (42).
Statistical Analysis-All experiments were performed in triplicate and repeated at least three times. Data are presented as percent changes compared with the controls ϮS.D. from the three independent experiments. Group means were compared by one-way ANOVA (Tukey's post hoc) using SPSS software, and a difference was accepted at p Ͻ 0.05 (39). The survival test was represented by Kaplan-Meier survival curves using SPSS software.

RESULTS
Decreased Survival in cav1 KO Mice-To determine the role of Cav-1 in P. aeruginosa infection, we intranasally instilled PAK (0.5 ϫ 10 7 cfu/mouse) to cav1 KO and WT mice (with otherwise similar genetic backgrounds). As shown in Fig. 1, cav1 KO mice exhibited increased lethality (50% cav1 KO mice died within 24 h postinfection). At 60 h, all KO mice died, whereas 50% of the WT control mice remained alive. The result is represented by Kaplan-Meier survival curves (p ϭ 0.017, log rank test). These findings suggest that Cav-1 is required for host defense in P. aeruginosa infection in acute pneumonia models.
Increased Bacterial Burdens and Oxidation in cav1 KO Lungs-To fully analyze the cause of the lethality, moribund mice were killed, and organs were aseptically removed for various assays. Lung homogenates were used to measure bacterial burdens. Cav-1 KO mice showed significantly increased colony-forming units of PAK in the lung tissue and AM cells when compared with WT mice ( Fig. 2A; p Ͻ 0.001, one-way ANOVA), indicating severe lung injury and pneumonia. To quantitatively determine polymorphonuclear neutrophil (PMN) infiltration, BAL fluid and serum were analyzed for the percentage of PMN cells. PMN penetration to the lung was increased in the BAL fluid and serum of cav1 KO mice when compared with WT mice (Fig. 2, B and C). P. aeruginosa infection was previously shown to induce the release of reactive oxygen species (ROS) in lung, which may accumulate and eventually cause lung injury (43). To measure this oxidative stress, cav1 KO and WT mice were infected by PAK, and AM cells were obtained. Compared with WT mice, AM cells of cav1 KO mice showed increased oxidative stress 18 h postinfection as determined by NBT assays (Fig. 2D). These results were further confirmed using the dihydrodichlorofluorescein diacetate assay (Fig. 2E), a sensitive fluorescence method for quantifying superoxide. Moreover, a decreased mitochondrial membrane potential was also observed in cav1 KO mice as compared with WT mice using JC-1 fluorescence assays (Fig. 2F), indicating that increased oxidation resulted in apoptotic cell death. These data collectively suggest that increased ROS may hamper cell survival by increasing apoptosis and may significantly damage the lung and other organs. These results are consistent with the increased PMN penetration in the lung and serum in cav1 KO mice (Fig. 2, B and C).
Under P. aeruginosa infection, PMNs are expected to infiltrate into the lung to clear bacteria through ROS and proteases. Although bacterial clearance is dependent in part on the production of ROS, excessive ROS accumulation may cause lung injury. As a direct indicator of lung injury, lung histology was examined 18 h post-PAK infection. Although both cav1 KO mice and WT control mice showed signs of pneumonia, significant histological alterations and increased PMN infiltration were only observed in the lungs of cav1 KO mice, indicating more severe lung injury in these animals (Fig. 2G). The insets demonstrate the typical regions with serious tissue damage or inflammatory response.
Infection Dissemination in cav1 KO Mice-A logical question is whether or not the intranasal inoculation only caused a local infection within the lung. Using tissue homogenates, we assessed bacterial burdens in several other organs. Colonyforming units of PAK were also increased in the spleen, liver, and kidneys, indicating that bacteria were able to spread from the original inoculation site (lungs) (Fig. 3, A-C). The spleen, which is the most sensitive indicator for bacterial spread, displayed a marked increase in bacterial burdens (Fig. 3A). Consistent with the PMN penetration into the serum (Fig. 2C), these results also suggest that the severe lung injury and bacterial spread into other organs are the possible cause of mortality in these mice.
To confirm the dissemination results, we next determined MPO activity in the lung and other organs to again probe neutrophil penetration because MPO is a recognized influx for oxidation in tissue. As expected, similar MPO activity was observed in the lung, liver, and kidney (Fig. 3, D-F). Increased MPO in organs other than the lung (e.g. liver and kidney) suggests that the superoxide release may result from systemic spread of the invading PAK bacteria.
As increased oxidation can also cause tissue injury by oxidative degradation of lipids, we used a thiobarbituric acid-reactive substance assay to detect lipid peroxidation in the lung, liver, and kidney tissue. Our results show that lipid peroxidation was significantly increased in all of the PAK-infected organs of cav1 KO mice as compared with WT mice (Fig. 3, G-I). The lung showed a marked increase in lipid peroxidation compared with the liver and kidney, suggesting that the lung was the main target. These data were consistent with the severity of lung injury determined using the MPO assay and superoxide detection, suggesting that the increased lipid peroxidation in KO mice is relevant to the progression of lung injury by oxidation.
Altered AM Function in cav1 KO Mice-AM cells play a crucial role in bacterial clearance by phagocytosis (43). To explore

Cav-1 Deficiency Impacts P. aeruginosa Infection
this relationship, we also measured the ingested PAK bacteria in AM cells. Increased bacterial burdens were found in the AM cells of cav1 KO mice compared with those of WT mice as assessed by colony-forming units (Fig. 4A). However, this result still cannot address whether the bacteria were dealt with through active phagocytosis or passive uptake. To determine the function of AM cells, equal amounts of PAO1-GFP, a strain of P. aeruginosa with green fluorescence (31,39), were added into cultured AM cells isolated from cav1 KO mice or WT mice for a 2-h incubation. This allowed us to conveniently quantify phagocytosis by measuring fluorescence intensity using a fluorescence plate reader (BioTek) (29,32). Phagocytosis of cav1 KO AM cells was significantly lower than that of WT AM cells, indicating that Cav-1 may play a role in phagocytosis by AM cells (Fig. 4B). These results also indicate that increased preingestion of the pathogen resulted in reduced phagocytic capacity in AM cells following infection.
Survival of AM cells postinfection is also an essential factor for maintaining their bacterial clearance function. Thus, to determine the AM viability, AM cells were infected with PAK for 1 h at m.o.i. 10:1. After adding polymyxin B (to remove the surface bacteria; Sigma-Aldrich), survival levels of AM cells were measured, which again showed a decrease in the survival of AM cells in cav1 KO mice after PAK infection (Fig. 4C).
These results indicate that Cav-1 may play a critical role in AM-mediated host defense against PAK infection.
Altered Inflammatory Response in cav1 KO Mice-Cav-1 has been implicated in bacterially induced inflammatory responses. A, the lungs showed significantly increased bacterial burdens after infection with PAK in cav1 KO mice compared with WT mice. cav1 KO mice and WT mice were infected with 0.5 ϫ 10 7 cfu/mouse PAK. Frozen tissues were homogenized in PBS. The same amounts of tissue were evaluated for bacterial colonies for cfu/g of tissue. B and C, increased PMN infiltration and an acute inflammatory response were observed in the lung and serum of cav1 KO mice compared with WT mice. After HEMA-3 staining (Fisher Scientific), PMN cell percentages were calculated versus total nuclear cells. D, superoxide production of AM cells was significantly increased in cav1 KO mice compared with WT mice using NBT assay (1 g/ml). The same amounts of AM cells were infected with PAK at m.o.i. 10:1 for 1 h. The optical density for NBT was quantified at 560 nm. E, oxidative stress was increased in cav1 KO mice as determined by the dihydrodichlorofluorescein diacetate (H2DCF) assay (5 M). The fluorescence was quantified at 488 nm. F, cav1 KO mice showed significantly lower mitochondrial potential than WT mice as assessed by the JC-1 fluorescence assay (0.1 g/ml). The fluorescence was quantified at 532 nm. G, increased lung injury and inflammation as assessed by histology. cav1 KO mice and WT mice were infected with 0.5 ϫ 10 7 cfu/mouse PAK for 24 h. Mice were dissected, and lungs were embedded in formalin. Sections were analyzed by H&E staining. The data are representative of three experiments (one-way ANOVA (Tukey's post hoc); *, p Ͻ 0.05; **, p Ͻ 0.01). RLU, relative luciferase units;RFU, relative fluorescence units. Error bars represent standard deviations.

Cav-1 Deficiency Impacts P. aeruginosa Infection
Cytokine concentrations of BAL fluid were measured 18 h post-PAK infection to probe various cytokines for reflecting the inflammatory response. The BAL fluid of cav1 KO mice showed no significant difference in inflammatory response compared with that of WT mice. However, in cav1 mouse lungs, IL-2 and IL-10 levels were significantly elevated compared with WT mice as assayed by ELISA (p Ͻ 0.05). Furthermore, TNF-␣ and IL-6 levels were increased by ϳ4-fold in cav1 KO mice (p Ͻ 0.01) (Fig. 5, A-D). This indicates that the lung epithelium may be the main source of increased cytokines. To validate the above data, mRNA levels of the above quantified cytokines were also measured by semiquantitative RT-PCR. mRNA expressions of TNF-␣, IL-6, IFN-␥, and IL-12a were all up-regulated by ϳ2-4-fold as determined by densitometry (Fig. 5E). Similarly, levels of TNF-␣, IFN-␥, IL-1␤, IL-12a, and IL-6 (particularly IL-12a and IL-6) were increased in cav1 KO lung tissues as assessed by Western blotting with densitometrical analysis (Fig.  5F). Collectively, these data indicate that cav1 KO mice showed a more intense proinflammatory response (TNF-␣, IL-1␤, and IL-6) following PAK infection as compared with WT mice.
Activation of NF-B Pathway and STAT3 by PAK Infection in cav1 KO Mice-To illustrate the mechanism that causes the dysregulated response to PAK infection in cav1 KO mice, we assessed several signaling proteins in lung tissues by Western blotting (Fig. 5G). We found that NF-B was highly activated in protein expression (2-fold) and phosphorylation (1.4-fold). To

Cav-1 Deficiency Impacts P. aeruginosa Infection
further dissect the pathway of cytokine response, we also analyzed STAT activation. STAT3, a transcription factor associated with NF-B activation in various cellular processes, showed significant expression in cav1 KO mouse lungs (4-fold), and phospho-STAT3 was drastically increased (24-fold; Fig.  5G). Because STAT5 was not significantly increased (data not shown), our data suggest that STAT3 might play a specific regulatory role in this infection (20). We then evaluated the immediate upstream regulator of STAT3 and found that phospho-JAK2 was dramatically increased in cav1 KO mice (34-fold) (Fig. 5G). We also analyzed the roles of SOCS3, a negative regulator of STAT3. A significant increase in SOCS3 was seen in cav1 KO mice as compared with WT mice by Western blotting analysis (data not shown), suggesting that SOCS3 was responsive to the infection-induced inflammatory response. These results strongly indicate the activation of STAT3 in cav1 KO mice as compared with WT mice.
PAO1 Infection Displays Similar Phenotypes in cav1 KO Mice-To determine whether the infection response and its mechanism are similar across different bacterial strains, we studied the immune response to another well studied P. aeruginosa strain, PAO1, in cav1 KO mice. A similar infection condition was used except for a higher cfu count because PAO1 is less cytotoxic than PAK. Consistent with observations in PAK infection, cav1 KO mice also showed elevated bacterial burdens in lungs and in other major organs (the liver, spleen, and kidney) after PAO1 infection (Fig. 6, A-D). Our findings demonstrate that lung tissues were severely injured in cav1 KO mice as compared with WT mice. In addition, sepsis was also observed in cav1 KO mice versus WT mice by PAO1 infection. Further-more, the inflammatory response indicated by cytokines and superoxide was also increased in cav1 KO mice following PAO1 infection (data not shown). We then investigated the NF-B pathway and STAT3 activation using Western blotting analysis. Similar to the PAK strain, the PAO1 strain also induced significant inflammatory responses (phospho-NF-B, 6.4-fold; phospho-STAT3, 5.4-fold; and phospho-JAK2, 5.8-fold) in cav1 KO mice as compared with WT mice. These results indicate that PAO1 infection also activated the STAT3 pathway along with NF-B activation (Fig. 6E). In addition, levels of inflammatory cytokines IL-6 (7.3-fold) and TNF-␣ (5.6-fold) were also increased in cav1 KO mice as compared with WT mice. Interestingly, the SOCS3 level was much higher in cav1 KO mice than WT mice (data not shown) and was also more significantly activated than in the PAK-infected mice (Fig. 5G). These data indicate that PAO1 infection in cav1 KO mice exhibited a phenotype similar to that of the PAK strain and that a more rigorous inflammatory response was induced by the lower cytotoxicity strain PAO1 as compared with PAK.

NF-B Pathway Is Altered by Mutating Cav-1 in MLE-12 Cells and by STAT3
Inhibitor-To better understand the role of STAT3 in the NF-B pathway, we transfected murine alveolar epithelial MLE-12 cells with either WT cav1-or a DN cav1expressing plasmid (45). MLE-12 cells have been widely used as a model for analyzing murine lung epithelial function (39,41,46). Transfected cells were then infected with PAK for 1 h at m.o.i. 10:1 and lysed for the cfu assay. As expected, increased bacterial burdens were observed in cav1 DN MLE-12 cells (Fig.  7A). Next, we assessed the mRNA levels for NF-B, JAK2, and STAT3, which were not significantly activated at an early phase (1 h postinfection). However, mRNA levels of IL-6 and IL-12a significantly increased. To ascertain the role of STAT3, WP1066 (a STAT3 inhibitor; Sigma-Aldrich) was used to pre-treat the cav1 DN cells. WP1066 has been demonstrated to inhibit the phosphorylation of STAT3 at the low M range, thereby effectively blocking STAT3 signaling (47). After treat-  ment with the STAT3 inhibitor, activation of IL-6 and TNF-␣ was abolished, whereas the total mRNA levels were not altered (Fig. 7B). These data support the view that STAT3 plays a crucial role in cytokine production. Furthermore, we determined the protein levels of the aforementioned cytokines and their regulators. Consistent with the mRNA expression, the protein levels of IL-6 and TNF-␣ were increased in cav1 DN-transfected cells when compared with WT controls by semiquantitative RT-PCR (Fig. 7C), which showed that cav1 DN transfection resulted in a profile similar to that observed in cav1 KO mice. By contrast, protein levels of IL-6 and TNF-␣ were suppressed by WP1066 pretreatment. Importantly, phosphorylation of NF-B, JAK2, and STAT3 was significantly increased in cav1 DN-transfected cells versus the control. Similarly, WP1066 pretreatment abolished the phosphorylation of NF-B, JAK2, and STAT3 in cav1 DN cells (Fig. 7C). These results confirm our findings observed in cav1 KO mice, indicating that the typical phenotypes may be attributable to a dysregulated proinflammatory cytokine response through the Cav-1/STAT3/NF-B axis.
Inactive (resting) NF-B and STAT3 are normally retained in the cytosol of MLE-12 cells (48). We hypothesized that these factors can be activated by infection, which may induce their translocation to the nucleus (49). We examined the localization of NF-B by confocal microscopy and found that both NF-B and STAT3 were activated and translocated into the nucleus in cav1 DN MLE-12 cells but not in vector-treated MLE-12 cells (Fig. 8, A and B). To further probe the activation of NF-B, we transfected MLE-12 cells and determined luciferase expression. Consistent with the imaging data, the luciferase promoter assay defined the activation of NF-B (Fig. 8C). Importantly, we reaffirmed that the activation of NF-B was also associated with the STAT3 pathway because the elevation in luciferase activity was significantly inhibited by WP1066. A simplified model represents the pathway implicated in cav1 KO mice and cav1 DN cells (Fig. 8D).

DISCUSSION
In this study, we investigated the phenotype of P. aeruginosa infection in cav1 KO mice. Severe disease phenotypes in cav1 KO mice were observed in our experiments, including decreased survival, increased inflammatory response, and more severe lung injury compared with WT mice. We reasoned that dysregulated cytokine response and increased superoxide release contribute to this exaggerated pathogenesis. Importantly, we have identified a novel mechanism involved in the phenotype in which the STAT3 pathway is responsible for the dysregulated inflammatory response. Our results are consistent with the hypothesis that Cav-1 plays an essential role in the clearance of P. aeruginosa because both phagocytosis of the bacteria by alveolar macrophages and cytokine production by alveolar epithelial cells were Cav-1-dependent (50). Similar phenotypes (increased susceptibility, increased production of cytokines, and elevated bacterial burdens) have been observed in cav1 KO mice infected with other pathogens, such as Salmonella typhimurium (51). Thus, Cav-1 may render critical immune defense against a wide range of microorganisms.
Previous studies have demonstrated that caveolins are widely expressed in lung epithelial cells (52,53). Because caveolins are associated with lipid rafts (54), they are thought to be an important part of the host response to P. aeruginosa infection in lung epithelial cells (7). The majority of previous studies have focused on illustrating the role of caveolins in regulating the cell responses to pathogen virulence factors like LPS (28,(55)(56)(57). However, limited reports have actually investigated the role of Cav-1 in the innate immunity of the lung against P. aeruginosa infection. Two previous reports showed different phenotypes in cav1 KO mice with a single pathogen, P. aeruginosa (11,12). This discrepancy must be addressed to understand the real role of this critical protein. Our timely studies reveal that Cav-1 is indispensable in host defense against P. aeruginosa infection, consistent with the observation made by Gadjeva et al. (12). Furthermore, our studies reveal mechanistically that Cav-1 is a powerful negative regulator of proinflammatory cytokines during this infection.
We have demonstrated that ROS levels were significantly increased and that mitochondrial potential was decreased in AM cells of cav1 KO mice as compared with those in WT mice. These data showed an elevated oxidative stress in cav1 KO mice that might cause lung injury and systemic bacterial infection. Overproduction of superoxide can impair innate immunity, including the phagocytosis function of AM cells, which form the front line of innate defense in the lung. Indeed, AM cells have been shown to be dysfunctional in cav1 KO mice, which exhibit reduced phagocytosis and increased susceptibility to cell death after P. aeruginosa infection. Although similar methods were utilized in a cross-pathogen comparison study by Lisanti and co-workers (51), no difference in bacterial ingestion was found between cav1 KO mice and WT mice with S. typhimurium infection. This discrepancy may be due to inherent differences in the two pathways or the distinct experimental settings. In addition, PMN cells, another important cell type responsible for bacterial clearance, show lowered phagocytosis against P. aeruginosa under Cav-1 deficiency (12). Together, these observations demonstrate that the bacterial clearance mechanism is impaired in Cav-1-deficient mice at least in the case of P. aeruginosa infection. Furthermore, the sustained recruitment of PMN cells into the lung necessary for bacterial clearance may release excess superoxide and proteases, resulting in more severe lung injury and systemic bacterial infection.
Another major contributing factor to lung injury and mortality is the inflammatory response. We found that the lungs of cav1 KO mice exhibit increases in proinflammatory cytokines, but a similar increase was not found in the BAL fluid of these mice. Thus, lung epithelial cells might also contribute to the production of cytokines (41,58,59), whereas macrophages and neutrophils are traditionally regarded as the main players in inflammatory responses (60). To further explore the dysregulation of inflammatory response in KO mice, modifications of the cell signaling pathway in the lungs were examined. In particular, we found NF-B, a major transcription factor for cytokine production in alveolar epithelial cells, to be highly activated (61).
To investigate the molecular events associated with this activation, we studied the JAK/STAT pathway. Interestingly, STAT3, but not STAT5, was found to be highly activated in cav1 KO mice following infection. The interaction between NF-B and STAT3 has been discussed extensively in various experimental models, particularly in cancer research, whereas the STAT3/NF-B axis is still relatively undefined in Gramnegative infections. Among the cytokines whose expression was altered, TNF-␣ and IL-6 were the most significantly upregulated. Considering that TNF-␣ and IL-6 are able to induce the activation of NF-B and STAT3, respectively, these significant changes might be partially attributable to positive feedback (22,62). Interestingly, we also observed significantly increased expression and phosphorylation of Akt and GSK3␤ (data not shown), which may be the other pathways responding to or resulting from the dysregulated cytokine production and may interact with the JAK2/STAT3 pathway in Cav-1-deficient animals. Consistent with our observations, previous reports have shown the relevance of the PI3K/Akt axis (27) and GSK3␤ to a JAK/STAT pathway (63). These indicate that in cav1 KO mice infection by various bacteria may activate several different pathways to compensate for the loss of this critical protein.
Indeed, there is little doubt that Cav-1 is important for regulating lipid raft-mediated endocytosis and various other cellular events (32,64,65). Despite such activation in various pathways, the loss of Cav-1 cannot be completely rescued to combat infections and usually leads to lethal consequences because crucial cellular signaling systems may have been dysregulated. For example, Cav-1 has been reported previously to interact with STAT3 directly through molecular interactions (51). Deficiency in Cav-1 may impact STAT5 activity (44), whereas the ultimate effect on STAT5 activity is cell-dependent (51). Furthermore, no previous studies have linked Cav-1 with the STAT3 pathway in response to a respiratory pathogen. Thus, linking Cav-1 with STAT3/NF-B is a novel finding for this host response. Finally, the role of the STAT3 pathway in AM cells is currently unclear and is worth studying.
Thus, we hypothesized that STAT3 might play a vital role in regulating NF-B activation. We used MLE-12 cells to establish a Cav-1-deficient lung epithelial culture model. Our data indicate that cav1 DN plasmid-transfected MLE-12 cells showed a similar inflammatory response and cell signaling activation as seen in cav1 KO mice. STAT3 inhibitor WP1066 was then used to confirm the role of STAT3 in MLE-12 cells. As expected, STAT3 inhibition decreased NF-B activation and reduced the inflammatory response, indicating that STAT3 activation is essential for NF-B activation and downstream inflammatory response modifications. Taken together, our data indicate that Cav-1 initiates a negative regulation in the STAT3/NF-B pathway. Although SOCS3, a chief negative regulator of STAT3, was also increased in cav1 KO mice (data not shown), the overactivation of STAT3 and the associated inflammatory response were still not constrained. Thus, our study suggests that Cav-1 may potentiate the negative regulatory effects of SOCS3 on the inflammatory response, whereas loss of Cav-1 significantly dampens this role, making this an interesting question for future studies.
In conclusion, we have demonstrated a typical phenotype of P. aeruginosa infection in cav1 KO mice, and our data indicate an important role for Cav-1 in innate immunity in mice. Cav-1 deficiency impaired the phagocytic ability and other immune defense mechanisms, resulting in sustained infiltration of PMN cells into the lung and an intense inflammatory response. Along with the bacterial infection-induced oxidative stress, the accumulation of PMN cells and failure of bacterial clearance may cause severe lung injury and systemic bacterial dissemination, which finally results in host mortality. In addition, we revealed a novel Cav-1/STAT3/NF-B axis directly contributing to a dysregulated cytokine profile in cav1 KO mice. More importantly, we confirmed this Cav-1/STAT3/NF-B pathway in cav1 knockdown MLE-12 cells during P. aeruginosa infection. Taken together, these observations have provided an insight into the role of Cav-1 in innate immunity against P. aeruginosa and might indicate novel targets for clinical therapy.